Plug and Play Prior Regularized Algorithm for Acoustic Resolution Photoacoustic Microscopy Bioimaging System Enhancement.

As a novel biomedical imaging setup, acoustic resolution photoacoustic microscopy (AR-PAM) employs photoacoustic effect for visualizing vasculature and bio-structure. Nevertheless, this modality suffers from low imaging resolution which greatly limited the application for clinical usage. Model based and data driven algorithms have been developed to enhance the imaging resolution and quality, however, there exists limitations in both kinds of algorithms respectively. In this paper, we proposed a Plug and Play prior regularized AR-PAM imaging enhancement algorithm, which combined the merits of both model based and data driven algorithms, where the prior term is learned by a denoising network and served as sub-module to the optimization framework. The physical model is also derived based on acoustic beamforming theory that characterizes the AR-PAM degradation mechanism and directly incorporated into the algorithm. Finally, the proposed algorithm is applied to enhance the simulated and in vivo AR-PAM imaging results to prove its performance. Perceptually, the proposed algorithm achieves the best enhancement quality when compared with total variation based algorithm. Quantitatively, the evaluation metrics also achieve the best values for simulated and in vivo enhancement results.